A. Overview
Total organic carbon is a well-established water quality parameter that quantifies the overall concentration of organic substances, all of which are typically regarded as contaminants, in an aqueous environment. Total organic carbon in an aqueous sample may be composed of either one or two components—dissolved organic carbon (DOC) and particulate organic carbon. The measurement of DOC is conventionally accomplished by filtering the water sample, commonly through a 0.45-μm filter, to remove particulate organic carbon prior to performing an analysis for DOC. The limitations of conventional apparatus and techniques for such analysis often lead to the result that only DOC is effectively measured, instead of TOC, because the particulates in a sample containing both forms of organic carbon typically cause errors in the measurement and plug fluid passages causing hardware failures.
In the following description, ‘DOC’ is used to refer to measurements in which the sample has first been filtered to remove particulates, while ‘TOC’ is used herein to refer to measurements in which the sample has not been filtered. In other respects, however, the following description is relevant to both DOC and TOC measurements.
In one known approach DOC and/or TOC, the organic compounds in an aqueous sample are oxidized to carbon dioxide (CO2) and the CO2 in the sample is then measured. In addition to organic carbon components, the water sample may initially contain CO2 and other inorganic forms of carbon (e.g., in the form of bicarbonate and carbonate salts). Together, these forms of inorganic carbon are referred to herein as IC. Total carbon (TC) concentration in an aqueous sample is therefore the sum of the TOC and IC concentrations.
Because an aqueous sample following an oxidation step could contain CO2 originating from both IC and TOC sources, the IC must be accounted for in some way to accurately measure TOC. One way to deal with IC is to remove the IC from the sample before the sample is oxidized. This is commonly done by acidifying the sample to convert carbonates and bicarbonates into free CO2, and then sparging it with CO2-free gas to remove that CO2 originating from IC sources. It has been found, however, that at least some volatile organic compounds may be removed from the sample during such a sparging step. Thus, when the sparged sample is subsequently oxidized, the CO2 produced is from the oxidation of the remaining (non-purgeable) organics in the sample, so this measurement is often referred to as the measurement of non-purgeable organic carbon (NPOC). Since many samples contain few, if any, purgeable organic compounds, the concentration of NPOC in those samples is typically considered to be essentially equal to the TOC concentration.
A second way of dealing with IC in a sample is to separately measure the TC and IC concentrations. When using this approach, the TOC concentration is calculated from the concentration difference, TC minus IC (TC−IC). One advantage of this approach is that the sample is not sparged so that purgeable organics are not lost thereby eliminating this source of measurement errors. As a result, the measurement of TOC by this “difference” approach is potentially more accurate.
When the approach to measuring carbon concentrations in aqueous samples is not constrained by requirements for regulatory compliance, a technician usually selects the parameter to be measured based on the time and resources required for each measurement, and the expected composition of the samples being monitored. Often, NPOC measurements are performed because they are relatively fast. Where it is necessary to accommodate a variable IC concentration, or where loss of purgeable organic carbon results in too large a discrepancy to be tolerated, TOC (or DOC) is measured by the difference approach, as described above.
In other cases, the samples may either contain IC concentrations that are known to be small compared to the TOC concentration, or the IC concentrations are relatively constant. In such cases, a technician may elect to measure TC because it is fast and it provides a sufficiently accurate indication of TOC trends for many common applications.
B. Identification of Related Prior Art
The following U.S. patents, each of which is incorporated herein by reference, are representative of pertinent prior art patents in the field of this invention and in related technical areas, such as oxidation of organic wastes: U.S. Pat. Nos. 3,296,435 (Teal '435); 3,700,891 (Luft '891); 3,958,945 (Takahashi '945); 4,619,902 (Bernard '902); 4,882,098 (Weetman '098); 4,896,971 (Weetman '971); 4,902,896 (Fertig '896); 5,037,067 (Ray '067); 5,232,604 (Swallow '604); 5,271,900 (Morita '900); 5,482,077 (Serafin '077); 5,630,444 (Callaghan '444); 5,835,216 (Koskinen '216); 6,007,777 (Purcell '777); 6,114,700 (Blades '700); 6,142,458 (Howk '458); 6,375,900 B1 (Lee-Alvarez '900); and 6,988,825 B2 (Colville '825).
The following technical publications, which are also incorporated herein by reference, are also representative of the pertinent prior art in this field: Aiken, G. R., “Chloride Interference in the Analysis of Dissolved Organic Carbon by the Wet Oxidation Method,” Environ. Sci. Technol., Vol. 26, No. 12, pp. 2435-2439; 1992; le Clercq, M.; van der Plicht, J. and Meijer, H. A. J., “A Supercritical Oxidation System for the Determination of Carbon Isotope Ratios in Marine Dissolved Organic Carbon,” Analyt. Chim. Acta, Vol. 370(1), pp. 19-27; 1998; Eyerer, P., “TOC Measurements on the Basis of Supercritical Water Oxidation,” AE-2e.1; Fraunhover-Gesellschaft zur Foerderung, Institut Chemische Technologie; Munich, Germany, http://www.ict.fraunhofer.de/english/projects/meas/onlan/index.html#a5; ISO-CEN EN 1484, “Water Analysis Guidelines for the Determination of Total Organic Carbon (TOC) and Dissolved Organic Carbon (DOC),” 1997; Koprivnjak, J-F, et al., “The Underestimation of Concentrations of Dissolved Organic Carbon in Freshwaters,” Water Research, Vol. 29, No. 1, pp. 91-94, 1995; Menzel, D. W. and Vaccaro, R. F., “The Measurement of Dissolved Organic and Particulate Carbon in Seawater,” Limnology and Oceanography, Vol. 9(1), pp. 138-142, 1964; Nitta, M.; Iwata, T.; Sanui, Y. and Ogawachi, T., “Determination of Total Organic Carbon in Highly Purified Water by Wet Oxidation at High Temperature and High Pressure,” presented at Tenth Annual Semiconductor Pure Water Conference; Feb. 26-28, 1991, Santa Clara, Calif.; in Conference Proceedings, Balazs, M. K. (Ed.), pp. 314-320; Wangersky, P. J., “Dissolved Organic Carbon Methods: A Critical Review,” Marine Chem., Vol. 41, pp. 61-74, 1993; and, William, P. J. leB. et al., “DOC Subgroup Report,” Marine Chem., Vol. 41, pp. 11-21, 1993. These patents and technical publications are further referred to in the following description.
C. Prior Art Related to Sample Handling and Sparging
Many prior art analyzers, such as those described in Morita '900, Purcell '777 and Lee-Alvarez '900, draw the sample into a syringe pump. Those syringe pumps use rotary valves to connect the syringe to the sample, reagents, dilution water and other analyzer apparatus. Any salts and particulates in the sample contact the sealing surfaces of the valve and syringe. As particles settle onto those surfaces, they cause wear and premature leaking. Salts also dry on the surfaces and cause additional wear because the salt crystals are abrasive.
Efficient sparging has been a goal of certain of the prior art patents and publications. Takahashi '945, for example, describes a multi-stage sparger for use in TOC analyzers. Employing more than one stage improves the efficiency of the sparging process. However, the large internal volume of this device would make it difficult to flush out between uses with different samples that have widely different concentrations of contaminants.
Weetman '098, Weetman '971, and Howk '458 teach that sparging can be made more efficient by stirring the solution with rotating propellers. This modified sparging method would be complex to incorporate in an analytical instrument, however, because of the motor and rotating seals that would be required.
In some prior carbon analyzers that add reagents to the sample, mixing is facilitated by bubbling gas through the solution (for example, Purcell '777). As discussed above, however, some samples should be measured without sparging because sparging can remove volatile organics and thereby introduce an error into the measurements. However, no analyzer is known to incorporate a device that can be used to sparge certain samples when desired, while also mixing other samples with reagents without sparging. Further, in prior analyzers, when the sparging stops, any particulates in the solution settle to the bottom of the sparger. This makes it impossible to accurately measure organic material in the particles, and it increases the likelihood that fluid passages of the apparatus will become plugged.
D. Prior Art Related to Oxidation Techniques
It is well known in the art to oxidize organic carbon using wet chemical methods. For example, in the Menzel and Vacarro publication, the authors report measuring DOC and particulate organic carbon in seawater by oxidizing a 5 mL sample in a sealed glass ampoule that also contained the oxidizing agent potassium persulfate. The oxidation was achieved by heating the ampoule to 130° C. for 30 min. After the heating step, the ampoule was cooled, broken open, and the CO2 contained inside it was measured using a non-dispersive infrared (NDIR) detector. Among other disadvantages, this method has the disadvantage that it involves many manual steps. Furthermore, the ampoules can break when they are heated or handled, raising concerns about loss of data and safety. This method would be impractical for real-time monitoring of process streams, or even laboratory analyses where many samples are to be analyzed each day.
Bernard '902 describes an instrument that automates the wet chemical oxidation method. The sample is acidified and a persulfate-containing reagent is added prior to the oxidation. CO2-free gas is bubbled through the sample to remove the IC (in preparation for making an NPOC measurement) or to transfer it to a NDIR detector for measurement of the IC. The solution is then heated to 90 to 100° C. at ambient pressure to achieve oxidation of the organics. During the oxidation, the CO2 is transferred to the NDIR detector where it is measured. The oxidation by persulfate at these temperatures is slow; in fact, the innovative aspect of Bernard '902 is the use of metal catalysts to increase the rate of the oxidation.
Another shortcoming of such wet chemical methods is that the oxidation of organics by prior wet chemical methods is incomplete, especially when the sample contains chloride [as reported for example in the publications of William, et al.; Wangersky; Koprivnjak, et al.; and Aiken]. When the oxidation is incomplete, the TOC measurement is inaccurate because not all of the organic carbon is measured.
Purcell '777 describes another analytical instrument that measures carbon in aqueous samples. In this case, the sample is acidified and an oxidizing reagent (a solution containing persulfate salts) is added to the sample. This mixing occurs in a syringe, and the resulting solution is then transferred to a sparger. After sparging, the syringe transfers the sample to a reactor where the solution is irradiated with ultraviolet (UV) radiation. In the presence of the UV radiation and the persulfate reagent, many organics in the sample are oxidized to CO2, and the CO2 is measured in a NDIR detector.
A problem with the oxidation of organics using UV radiation, as in Purcell '777, is that it is inefficient when the sample contains particulates. For example, one study reported that TOC analyzers that are based on UV oxidation detected less than 3% of the cellulose particles added to samples at an actual concentration of 20 mg C/L. By comparison, analyzers based on high-temperature catalytic oxidation (HTCO), as described below, on average detected 83.2% of the cellulose represented by cellulose particles.
To achieve more complete oxidation and, therefore, greater accuracy, analyzers were developed that oxidize organics using HTCO. Teal '435, for example, teaches that TOC in aqueous samples can be measured by injecting a portion of an IC-free sample into a catalytic reactor heated to around 900° C. The water vaporizes immediately, and organic materials are oxidized to CO2 upon contact with the catalyst. A carrier gas (oxygen) transports the CO2 out of the reactor to a NDIR detector.
Morita '900 and Lee-Alvarez '900 describe methods and apparatus that automatically acidify and sparge samples, oxidize organics using HTCO, and use NDIR detectors to measure the CO2. In both of these approaches, the sample is mixed with acid in a syringe connected to a multi-port valve that directs fluids to other components. Morita '900 describes the sparging as being performed inside the syringe, while Lee-Alvarez '900 describes a separate sparger.
A shortcoming of all methods based on HTCO is that samples containing salts or particulate material will eventually plug the reactor because the water evaporates in the reactor, leaving nonvolatile salts and particulates behind. In addition, the reactor typically requires two hours or more to cool enough so that it can be safely removed and cleaned. Then, about another two hours are required for the reactor to heat back up to its operating temperature. This means that the instrument is out of service for an extended period whenever the reactor must be cleaned.
Other oxidation methods have also been reported. The Nitta et al. technical publication describes an analyzer in which the sample is mixed with sulfuric acid and sodium persulfate. IC is removed by sparging, and then a pump pressurizes a continuously flowing stream of the solution to 2.0 to 2.5 MPa (284 to 356 psig). The pressurized solution is heated in a reactor to 200° C., and the organics are oxidized to CO2. The solution then flows through a flow restrictor (it is the flow through this restrictor that generates the upstream pressure as stated above). The CO2 produced during the oxidation is removed by sparging and is measured using an infrared detector. Several Japanese patents describe additional aspects of the instrument as described above (JP63135858, JP1021352, JP1021355, JP1021356, JP1021356, JP1049957, JP1049958, JP1318954, JP1318955, JP1318956, and JP5080022). This method has the advantage that higher oxidation temperatures, presumably with more complete oxidation, can be achieved than if the oxidation were performed at ambient pressure. However, the apparatus has the disadvantage that particles and salts will rapidly plug the restrictor. Furthermore, the cost of such an apparatus is likely to be high because solution has to be pumped continuously against the backpressure generated by the restrictor.
Attempts to improve the efficiency of the oxidation also have included oxidizing samples under supercritical conditions (i.e., above 374° C. and pressures above 22.12 MPa). Le Clercq et al. reported measuring carbon isotope ratios in DOC. Between 500 and 1,000 mL of seawater were mixed with oxygen, pumped to a maximum pressure of 35 MPa, and forced through an alumina reactor heated to 650° C. Placing a 0.18 mm ID capillary downstream of the reactor and setting the flow rate at 2 mL/min produced the aforementioned pressure. The gases exiting the capillary were cooled to collect the CO2 formed during the oxidation, and a mass spectrometer was used to measure the isotopes of carbon in the CO2. A problem with this apparatus is that samples that contain particulates tend to plug the capillary. Such a problem was reported by le Clercq et al., and they installed a 2-μm filter ahead of the capillary in an attempt to mitigate the problem. However, in applications in which the apparatus must operate for long periods, even the periodic plugging of such a filter would create excessive maintenance and downtime. Another problem is the extremely high temperature and pressure at which the oxidation is made to occur. Appropriate hardware for such operating conditions is costly, and it is likely to corrode rapidly. The reactor described in this technical literature was made from alumina to minimize corrosion, but the structural characteristics of alumina make it unreliable. For this reason, the alumina reactor had to be mounted inside a metal shield for safety.
Eyerer reported another approach. The sample is first pumped through an electrochemical cell that generates the oxidizing agent. Then the sample passes through a reactor heated up to 600° C. and through a valve that creates a backpressure up to 26 MPa. The sample is oxidized at those conditions and then passes over a hydrophobic membrane. Some of the CO2 diffuses through the membrane and is measured in a mass spectrometer. This apparatus suffers from the same types of corrosion, reliability, and cost shortcomings, however, that were discussed above for the le Clercq et al. approach.
Beyond applications in the measurement of organic carbon, as described above, rapid oxidation also has been a goal of developers of organic waste destruction systems. One way of achieving the desired rapid oxidation rates has been to perform the oxidation at near-critical and supercritical conditions. A variety of oxidizing agents have been employed, and one of the most economically attractive oxidizers is the oxygen in air. Swallow '604 teaches that if ozone, hydrogen peroxide, or salts containing persulfate, permanganate, nitrate, and other oxygen-containing anions are added to the liquid/air mixture, the oxidation rate is sufficiently rapid that the exothermic process operates without supplemental heating. This is an important consideration for large industrial processes, but it is much less important to analytical instrumentation because the hardware is much smaller.
Instead of oxidizing organics in a continuously flowing stream, batches of the sample could be heated. To do that requires that the batch be sealed in a container that is subsequently heated, and the best way of automatically sealing the container would be to use a valve that can withstand the pressure generated during heating. Many valves designed for high-pressure applications employ precision sealing surfaces. Ball valves require highly polished balls in packing glands to avoid leaks. Other high-pressure valves require metal-to-metal seals (for example, as described in Callaghan '444). Those valves are costly and subject to rapid wear by particles in the liquid.
A better method of achieving valve sealing in the presence of particulates is to use a softer seal that is resistant to abrasion. Serafin '077 teaches that elastomeric seals can be used in check valves at high pressure, and Ray '067 describes the use of O-rings to seal the ports in a plug valve. Both inventions have the disadvantage that the seals are not easily accessed for replacement when they become worn.
E. Prior Art Related to NDIR Detectors
NDIR detectors of CO2 used as components of TOC analyzers commonly use a rotating chopper wheel to modulate the infrared (IR) radiation, and a pneumatic IR detector to measure the IR radiation that has not been absorbed by the CO2 being measured. Luft '891 describes such a NDIR detector. Shortcomings of this technology include the fact that the chopper wheel mechanism is subject to failure, and irregularities in the size and orientation of the openings in the chopper wheel produce significant electrical noise in the measurement of CO2.
To overcome the effects of temperature and pressure on NDIR response, detectors with built-in temperature and pressure compensation have been reported, such as in Fertig '896. An alternative approach to overcoming temperature effects is to use an IR detector that is relatively unaffected by temperature, such as a pyroelectric detector. Koskinen '216 describes a NDIR detector that electronically modulates the IR source to avoid problems with chopper wheels, and it uses a pyroelectric IR detector. However, this NDIR uses a costly Fabry-Perot interferometer to select the IR wavelength that is measured.
Blades '700 describes a NDIR detector designed specifically for use in a TOC analyzer. The IR source is an electrically modulated incandescent lamp with a pyroelectric IR detector. However, the use of an incandescent lamp limits the dynamic range and sensitivity of the NDIR because the modulation is limited to low frequencies.
Commonly, NDIR detectors use rectifier circuits and lowpass filters to produce a DC signal that is proportional to the average output of the IR detector. Shortcomings of this technology include the conversion of “noise” over a wide bandwidth into a part of the rectifier output signal. Additionally, the lowpass filter that averages the rectified waveform also impairs the ability of the NDIR to respond to rapidly changing CO2 concentrations. Blades '700 reports an NDIR that uses two synchronous detectors, with each responding to opposite half-cycles of the signal from the IR detector. The use of two synchronous detectors improves the response time limitation of the rectifier circuit, but this approach still suffers from the shortcoming of mixing noise into the output signal.
Carbon measurement instruments commonly generate chlorine when the sample contains chloride ions. That chlorine would corrode many NDIR detectors, so scrubbers are used to remove the chlorine before it enters the NDIR (as, for example, in Lee-Alvarez '900 and Purcell '777). The scrubber is a consumable that adds to the operating cost and maintenance labor of those instruments.
These and other limitations of, and deficiencies in, the prior art approaches to IC, TOC and TC measurements are overcome in whole, or at least in part, by the methods and apparatus of this invention.